XPK Start: Sun Apr 19 19:36:42 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-19 19:37:06.697661: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0419 19:37:06.879807 136181852628800 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 19:37:15.920661 136181852628800 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 19:37:15.923170 136181852628800 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-rs-v8x8e-slice-job-0-0.mt-16-zarr3-true-rs-v8x8e:8482 I0419 19:37:17.303401 136181852628800 max_utils.py:284] Jax distributed system initialized! I0419 19:37:22.417880 136181852628800 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 19:37:22.417984 136181852628800 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 19:37:22.418025 136181852628800 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0419 19:37:22.418059 136181852628800 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 19:37:23.118433 136181852628800 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 19:37:23.118707 136181852628800 checkpointing.py:677] Setting up checkpoint logger... I0419 19:37:23.118761 136181852628800 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 19:37:23.118806 136181852628800 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 19:37:23.119141 136181852628800 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7bdab81ea510>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 19:37:24.779264 136181852628800 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0419 19:37:24.779551 136181852628800 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc6f0215c40>}, handler_registry=None I0419 19:37:24.779836 136181852628800 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc6f0215c40>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0419 19:37:24.779900 136181852628800 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc6f00b4b60>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0419 19:37:24.779947 136181852628800 composite_checkpoint_handler.py:505] Initialized registry DefaultCheckpointHandlerRegistry({('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc6f0215c40>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bc6f0215c40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc6f00b4b60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bc6f00b4b60>}). I0419 19:37:24.780308 136181852628800 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 19:37:24.780385 136181852628800 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7bc60042fd80> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 19:37:27.185917 136063179794176 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776627181483620081, 'commit_timestamp_nsecs': 1776627224106444647, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0419 19:37:27.432397 136063179794176 checkpoint_manager.py:1871] Missing metrics for step 0 I0419 19:37:28.278189 136181852628800 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints I0419 19:37:28.281013 136181852628800 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7bc6f021a1b0> I0419 19:37:28.281131 136181852628800 checkpointing.py:301] Checkpoint manager created! I0419 19:37:29.215688 136181852628800 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 19:37:29.215803 136181852628800 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 19:37:29.595773 136181852628800 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 19:37:29.595866 136181852628800 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 19:37:29.611880 136181852628800 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 19:37:29.611940 136181852628800 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 19:37:29.635355 136181852628800 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 19:37:29.635435 136181852628800 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 19:37:29.651630 136181852628800 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 19:37:29.651689 136181852628800 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 19:37:29.667807 136181852628800 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 19:37:29.667866 136181852628800 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 19:37:29.693081 136181852628800 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 19:37:29.693158 136181852628800 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 19:37:29.713905 136181852628800 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 19:37:29.713966 136181852628800 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 19:37:29.930057 136181852628800 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0419 19:37:29.930165 136181852628800 checkpointing.py:581] restoring from this run's directory step 9 I0419 19:37:30.174578 136181852628800 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/9. I0419 19:37:30.423578 136181852628800 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/9. I0419 19:37:30.424010 136181852628800 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/9. I0419 19:37:31.845702 1416 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 19:37:39.163444 136181852628800 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 215.773 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.323793411254883 s) (per-host) I0419 19:37:39.169110 136181852628800 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 204.590 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.735442638397217 s) (per-host) I0419 19:37:39.171366 136181852628800 event_tracking.py:138] [process=6] [sync] Finished load in 9.00 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/9 I0419 19:37:39.171480 136181852628800 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776627450.1744776, 'checkpointer_duration_secs': 8.996972799301147, 'checkpoint_manager_start_time': 1776627449.930432, 'checkpoint_manager_duration_secs': 9.241020917892456} fsdp: 32 I0419 19:37:39.586323 136181852628800 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0419 19:37:39.586467 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 19:37:39.586529 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 19:37:39.586590 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0419 19:37:39.586642 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 19:37:39.586679 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 19:37:39.586730 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0419 19:37:39.586783 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0419 19:37:39.586824 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0419 19:37:39.586860 136181852628800 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0419 19:37:39.586907 136181852628800 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0419 19:37:39.586955 136181852628800 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0419 19:37:40.076816 136181852628800 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 19:37:40.076913 136181852628800 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 19:37:40.092190 136181852628800 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 19:37:40.092252 136181852628800 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 19:37:51.017299 136181852628800 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0419 19:37:51.018150 136181852628800 metric_logger.py:301] number parameters: 1.104 billion I0419 19:38:02.405924 136181852628800 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0419 19:38:02.560721 136181852628800 checkpointing.py:776] Waited 0.15477871894836426 seconds for step 10 to finish before starting checkpointing. I0419 19:38:02.563364 136181852628800 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 19:38:02.565250 136181852628800 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10 I0419 19:38:02.566978 136181852628800 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/10. I0419 19:38:03.314309 136181852628800 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 19:38:03.315271 136181852628800 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 19:38:03.315329 136181852628800 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 19:38:03.593282 136181852628800 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.279051s I0419 19:38:03.593467 136181852628800 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.449 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.283113956451416 s) (per-host) I0419 19:38:03.593523 136181852628800 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.283181s (batch_requests_ready=0.002328s, total_serialization_initiated=0.280776s, others=0.000077s) I0419 19:38:03.593606 136181852628800 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.287307s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.287290) I0419 19:38:03.594277 136181852628800 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/10. I0419 19:38:03.594637 136062567421696 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:58:03.594600 I0419 19:38:03.602236 136181852628800 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 19:38:03.602489 136060952622848 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 19:38:03.602615 136181852628800 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776627482.563345, 'wait_for_prev_duration_secs': 8.177757263183594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776627482.5652876, 'checkpointer_blocking_duration_secs': 1.0294544696807861, 'get_old_steps_start_time': 1776627483.5947647, 'get_old_steps_duration_secs': 3.218650817871094e-05, 'checkpoint_manager_blocking_start_time': 1776627482.561232, 'checkpoint_manager_blocking_duration_secs': 1.0413479804992676} I0419 19:38:03.602719 136181852628800 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0419 19:38:03.602769 136181852628800 max_utils.py:750] Memstats: After params initialized: I0419 19:38:03.602816 136181852628800 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 19:38:03.602847 136181852628800 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 19:38:03.602874 136181852628800 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 19:38:03.602898 136181852628800 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 19:38:03.913650 136181852628800 metric_logger.py:196] completed step: 10, seconds: 11.388, TFLOP/s/device: 1.193, Tokens/s/device: 179.844, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0419 19:38:04.103477 136181852628800 metric_logger.py:196] completed step: 11, seconds: 1.506, TFLOP/s/device: 9.023, Tokens/s/device: 1359.990, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0419 19:38:04.251063 136181852628800 metric_logger.py:196] completed step: 12, seconds: 0.043, TFLOP/s/device: 318.244, Tokens/s/device: 47969.270, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0419 19:38:04.398542 136181852628800 metric_logger.py:196] completed step: 13, seconds: 0.154, TFLOP/s/device: 88.495, Tokens/s/device: 13338.978, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0419 19:38:04.403121 136181852628800 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0419 19:38:04.692787 136181852628800 checkpointing.py:776] Waited 0.289630651473999 seconds for step 15 to finish before starting checkpointing. I0419 19:38:04.695764 136181852628800 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 19:38:09.360043 136061493679872 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/10/items/array_metadatas/process_6 I0419 19:38:46.148170 136062567421696 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.874 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.83778190612793 s) (per-host) I0419 19:38:46.148298 136062567421696 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.553046s. I0419 19:38:54.849855 136062567421696 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.254585s. I0419 19:38:54.850147 136060952622848 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 19:38:54.850298 136060952622848 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 19:38:54.850366 136060952622848 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 19:38:54.851975 136060952622848 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 19:38:54.852160 136181852628800 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0419 19:38:54.852291 136181852628800 checkpoint_manager.py:1452] Waiting for previous save to complete took 50.156526 seconds. If this number is high, consider checkpointing less frequently. I0419 19:38:54.854283 136181852628800 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15 I0419 19:38:54.856754 136181852628800 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/15. I0419 19:38:55.568093 136181852628800 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 19:38:55.568192 136181852628800 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 19:38:55.609167 136181852628800 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042140s I0419 19:38:55.609334 136181852628800 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.741 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04571843147277832 s) (per-host) I0419 19:38:55.609383 136181852628800 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045777s (batch_requests_ready=0.001915s, total_serialization_initiated=0.043796s, others=0.000067s) I0419 19:38:55.609485 136181852628800 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050656s (all_items=0.000014s, per_item={'items': '0.00001383'}, temp_paths=0.050642) I0419 19:38:55.610125 136181852628800 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/15. I0419 19:38:55.610483 136061493679872 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:58:55.610446 I0419 19:38:55.612464 136181852628800 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 19:38:55.612767 136060952622848 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 19:38:55.612927 136181852628800 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776627484.6957345, 'wait_for_prev_duration_secs': 50.15652585029602, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776627534.8543212, 'checkpointer_blocking_duration_secs': 0.7563114166259766, 'get_old_steps_start_time': 1776627535.610655, 'get_old_steps_duration_secs': 3.218650817871094e-05, 'checkpoint_manager_blocking_start_time': 1776627484.6933715, 'checkpoint_manager_blocking_duration_secs': 50.9195237159729} I0419 19:38:55.613040 136181852628800 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0419 19:38:55.613868 136181852628800 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.058, Tokens/s/device: 13875.997, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0419 19:38:55.619880 136181852628800 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.187, Tokens/s/device: 13895.485, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0419 19:39:02.143175 136058286184192 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/15/items/array_metadatas/process_6 I0419 19:39:16.862249 136181852628800 metric_logger.py:196] completed step: 16, seconds: 51.215, TFLOP/s/device: 0.265, Tokens/s/device: 39.988, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0419 19:39:17.009516 136181852628800 metric_logger.py:196] completed step: 17, seconds: 21.096, TFLOP/s/device: 0.644, Tokens/s/device: 97.080, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0419 19:39:17.157081 136181852628800 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.133, Tokens/s/device: 13435.147, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0419 19:39:17.303804 136181852628800 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0419 19:39:17.304501 136181852628800 checkpointing.py:776] Waited 0.0007181167602539062 seconds for step 19 to finish before starting checkpointing. I0419 19:39:17.306599 136181852628800 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 19:39:38.087670 136061493679872 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.146 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.52401685714722 s) (per-host) I0419 19:39:38.087792 136061493679872 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.477192s. I0419 19:39:46.172768 136061493679872 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.562152s. I0419 19:39:46.173054 136060952622848 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 19:39:46.173171 136060952622848 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 19:39:46.173220 136060952622848 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 19:39:46.174620 136060952622848 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 19:39:46.174800 136181852628800 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0419 19:39:46.174923 136181852628800 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.868329 seconds. If this number is high, consider checkpointing less frequently. I0419 19:39:46.176423 136181852628800 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19 I0419 19:39:46.178433 136181852628800 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/19. I0419 19:39:46.883942 136181852628800 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 19:39:46.884036 136181852628800 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 19:39:46.924999 136181852628800 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041859s I0419 19:39:46.925151 136181852628800 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.939 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04545187950134277 s) (per-host) I0419 19:39:46.925199 136181852628800 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045510s (batch_requests_ready=0.001781s, total_serialization_initiated=0.043666s, others=0.000062s) I0419 19:39:46.925286 136181852628800 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.049744s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.049732) I0419 19:39:46.925988 136181852628800 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/19. I0419 19:39:46.926262 136060952622848 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:59:46.926229 I0419 19:39:46.928209 136181852628800 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 19:39:46.928506 136058810021632 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 19:39:46.928671 136181852628800 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776627557.30657, 'wait_for_prev_duration_secs': 28.86832904815674, 'time_between_consecutive_saves_sec': 22.454548120498657, 'checkpointer_blocking_start_time': 1776627586.1764705, 'checkpointer_blocking_duration_secs': 0.7499346733093262, 'get_old_steps_start_time': 1776627586.9264386, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776627557.3047304, 'checkpoint_manager_blocking_duration_secs': 29.623910665512085} I0419 19:39:46.928779 136181852628800 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0419 19:39:46.928822 136181852628800 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 19:39:52.652730 136058286184192 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/19/items/array_metadatas/process_6 I0419 19:40:28.848427 136060952622848 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.638 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.96867609024048 s) (per-host) I0419 19:40:28.848552 136060952622848 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.922175s. I0419 19:40:36.645498 136060952622848 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.719104s. I0419 19:40:36.645763 136058810021632 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 19:40:36.645904 136058810021632 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 19:40:36.645959 136058810021632 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 19:40:36.648426 136058810021632 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 19:40:36.649710 136181852628800 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0419 19:40:36.649851 136181852628800 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 19:40:36.650865 136181852628800 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.191, Tokens/s/device: 13896.051, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 19:40:47 UTC 2026 EXIT_CODE=0